Date of Award

Spring 2022

Project Type


Program or Major

Natural Resources and Environmental Studies

Degree Name

Doctor of Philosophy

First Advisor

Russell G Congalton

Second Advisor

Mark Ducey

Third Advisor

Meghan MacLean


The continued decline in forest cover across New England becomes more concerning when faced with the fact that these same forests may be playing an important role in the fight against climate change. New Hampshire, in particular, is experiencing a 0.27% annual net loss in forest cover as of 2018. Increased population growth and accompanied development has resulted in the removal of forest cover and the fragmentation of once continuous forest blocks. Fragmentation can lead to further degradation of the remaining forest stands via alterations of the biotic and abiotic process at their edges. The use of unmanned aerial systems (UAS) is becoming an important tool to ensure the sustainable management of current forests stands and may help to better understand the effects of fragmentation at forest edges. Because of the relatively recent arrival of this technology, effective and appropriate testing for accurate and efficient data collection is necessary. Furthermore, UAS have not been employed yet to detect edge effects.This research investigated the impacts of UAS flight parameters on the accuracy of canopy height estimates made from UAS data by comparing UAS estimates across twelve combinations of flying height and image overlap to ground measured canopy height. A multi-temporal approach to species level mapping with UAS imagery was tested by collecting multiple dates of UAS imagery from early spring to late summer and assessing whether the inclusion of one or more dates improved classification accuracy. Additional comparisons between RGB and multi-spectral cameras were carried out. Finally, UAS imagery was used to measure and assess the changes in canopy cover with increased distance from the edge. This trend was compared to trends in canopy cover measured on the ground. The results show that flying height had no impact of the accuracy of the height estimates made from UAS data and increasing forward image overlap resulted in a significant but minor increase in accuracy. Classification accuracy was improved with the use of multi-temporal data collection but no more than three dates of optimally timed imagery was necessary. Additionally, the RGB imagery produced maps with consistently higher accuracy than the multi-spectral sensor employed in this study. Finally, we were able to detect and measure a significant trend in canopy cover that mimicked the trends found on the ground. The results of the first two parts of this dissertation will go on to provide guidance to forestry practitioners on how to collect UAS that balances accuracy and efficiency, thus reducing project costs. The final result serves as an initial demonstration of utilizing UAS for understanding edge effects and opens the door to better understanding the impacts of fragmentation over larger areas.